Microsoft grounds its AI chat bot after it learns racism

It's not certain how Microsoft will teach Tay better manners, although it seems like word filters would be a good start. The company tells Business Insider that it's making "adjustments" to curb the AI's "inappropriate" remarks, so it's clearly aware that something has to change in its machine learning algorithms. Frankly, though, this kind of incident isn't a shock -- if we've learned anything in recent years, it's that leaving something completely open to input from the internet is guaranteed to invite abuse. Update: A Microsoft spokesperson has provided the statement that BI received. You can read the whole thing below.